Synthetic and mechanistic studies of the aza-retro-claisen rearrangement. a facile route to medium ring nitrogen heterocycles

Article abstract of DOI:10.1021/ol100397q

An efficient synthesis of medium-sized heterocyclic rings was achieved using a one-pot aza-Wittig/retro-aza-Claisen sequence of vinyl cyclobutanecarboxaldehydes derived from simple allylic carbonates. The use of various Staudinger reagents in the aza-Wittig reaction allows for a variety of N-substituted products to be obtained. The rearrangement is under thermodynamic control driven by relief of the cyclobutane ring strain and resonance stabilization of the resulting vinylogous amide/sulfonamide.

Full text of DOI:10.1021/ol100397q

ORGANIC
LETTERS
2010
Vol. 12, No. 7
1628-1631
Synthetic and Mechanistic Studies of
the Aza-Retro-Claisen Rearrangement.
A Facile Route to Medium Ring Nitrogen
Heterocycles
Robert K. Boeckman, Jr.,* Nathan E. Genung, Ke Chen, and Todd R. Ryder
Department of Chemistry, UniVersity of Rochester, Rochester, New York 14627-0216
rkb@rkbmac.chem.rochester.edu
Received February 16, 2010
ABSTRACT
An efficient synthesis of medium-sized heterocyclic rings was achieved using a one-pot aza-Wittig/retro-aza-Claisen sequence of vinyl
cyclobutanecarboxaldehydes derived from simple allylic carbonates. The use of various Staudinger reagents in the aza-Wittig reaction allows
for a variety of N-substituted products to be obtained. The rearrangement is under thermodynamic control driven by relief of the cyclobutane
ring strain and resonance stabilization of the resulting vinylogous amide/sulfonamide.
Heterocycles containing seven- or eight-membered rings are
common structural elements in an array of natural products,
many of which possess valuable biological activities.1 A
number of useful synthetic methods have been developed
for the generation of medium ring nitrogen heterocycles,2
including rearrangement of N-alkyl-2,3-divinyl and N-acyl-
3-divinylaziridines,3 2-imino-3-vinyl-cyclopropanes,4 and
ring-closing metathesis (RCM).5 The latter occasionally
encounters problems with functional group compatibility and
sluggish reaction rate.6 Methods to prepare partially saturated
eight-membered azacycles are much rarer.6,7 We sought to
extend our prior studies of the retro-Claisen reaction to
develop an efficient and general route to medium size
nitrogen-containing rings.8 Our prior studies that examined
the retro-Claisen rearrangement of vinylcyclopropane and
vinylcyclobutane carboxaldehydes afforded a useful entry
to highly substituted enantiomerically enriched oxepines and
oxacenes culminating in the total synthesis of (+)-laurenyne
(Figure 1).8d
(1) (a) Faulkner, D. J. Nat. Prod. Rep. 1984, 1, 251–280. (b) Faulkner,
D. J. Nat. Prod. Rep. 1986, 3, 1–33. (c) Tsuda, M.; Kobayashi, J.
Heterocycles 1997, 46, 765–794. (d) Matzanke, N.; Gregg, R.; Weinreb,
S. Org. Prep. Proced. Int. 1998, 30, 1–52. (e) Magnier, E.; Langlois, Y.
Tetrahedron 1998, 54, 6201–6258.
(2) (a) Hassenruck, K.; Martin, H. D. Synthesis 1988, 1988, 569–586.
(b) Shaw, R.; Anderson, M.; Gallagher, T. Synlett 1990, 1990, 584–586.
(c) Evans, P. A.; Holmes, A. B. Tetrahedron 1991, 47, 9131–9166. (d)
Furness, K.; Aube, J. Org. Lett. 1999, 1, 495–498. (e) Larock, R. C.; Tu,
C.; Pace, P. J. Org. Chem. 1998, 63, 6859–6866.
(6) Borer, B. C.; Deerenberg, S.; Bieraugel, H.; Pandit, U. K. Tetrahe-
dron Lett. 1994, 35, 3191–3194.
(7) (a) Chang, M.-Y.; Kung, Y.-H.; Ma, C.-C.; Chen, S.-T. Tetrahedron
2006, 63, 1339–1344. (b) Georg, G. I.; Guan, X. Tetrahedron Lett. 1992,
33, 17–20. Arata, Y.; Tanaka, K.; Yoshifuji, S.; Kanatomo, S. Chem. Pharm.
(3) (a) Lindstro¨m, U., M.; Somfai, P. Chem.sEur. J. 2001, 7, 94–98.
(b) Pommelet, J. C.; Chuche, J. Can. J. Chem. 1976, 54, 1571–1581.
(4) Reissig, H.-U.; Bo¨ttcher, G.; Zimmer, R. Can. J. Chem. 2004, 82,
166–176.
Bull. 1979, 27, 981–983.
(8) (a) Boeckman, R. K., Jr.; Flann, C. J.; Poss, K. M. J. Am. Chem.
Soc. 1985, 107, 4359–4362. (b) Boeckman, R. K., Jr.; Shair, M. D.; Vargas,
J. R.; Stolz, L. A. J. Org. Chem. 1993, 58, 1295–1297. (c) Boeckman, R. K.,
Jr.; Reeder, M. R. J. Org. Chem. 1997, 62, 6456–6457. (d) Boeckman,
R. K., Jr.; Zhang, J.; Reeder, M. R. Org. Lett. 2002, 4, 3891–3894.
(5) Grubbs, R. H.; Miller, S. J.; Fu, G. C. Acc. Chem. Res. 1995, 28,
446–452.
10.1021/ol100397q  2010 American Chemical Society
Published on Web 03/10/2010

thermodynamically favored equilibration to the azocine 8 under
the conditions required for conversion to 7 (Scheme 2).
Scheme 2. Potential One-Pot Aza-Retro-Claisen Rearrangement
Figure 1. Retro-Claisen rearrangement to form oxacycles.
In order to explore the means for introduction of nitrogen,
we first examined a stepwise process that was initiated by
reduction of racemic oxacene 1 to the derived alcohol 2
(Scheme 1).9 Upon heating, Claisen rearrangement was
observed to provide aldehyde 3 with no evidence of the retro-
Claisen isomer detectable by NMR (Keq >20) in spite of the
loss of vinylogous resonance stabilization by the electron-
withdrawing group.
Initially, the direct aza-retro-Claisen was attempted by
treating racemic oxacene 1 with excess benzyl amine (or
other primary amines) in the presence of MgSO4 (Scheme
3).11 Surprisingly, this led to preferential Michael addition-
elimination resulting in ring cleavage, affording the acyclic
hydroxy vinylogous formamide 9.
Scheme 3. Unexpected Ring Cleavage
Scheme 1. Stepwise Aza-Retro-Claisen Rearrangement
To obtain the desired regioselectivity during conversion
of 1 to imine 8, we sought to increase the nucleophilicity of
the amine via use of the related Staudinger imino-phospho-
ranes facilitating kinetic formation of the imine 8 via
elimination of triphenylphosphine oxide.12 Staudinger re-
agents are readily accessible via reaction of alkyl or aryl
azides with triphenylphosphine under mild heating.12
The N-benzyl Staudinger reagent (g2 equiv), generated
in situ from BnN3 and Ph3P in PhCH3 at reflux, was treated
with oxacene 1 until conversion to 11 was complete (∼14
h). Because the resulting azocine imine 11 proved to be unstable
to purification, we effected direct hydrolysis of the reaction
mixture with aq NaOAc buffer (pH ) 4) at rt, affording azocine
aldehyde 5 in 82% yield after chromatography.13
The precise sequence of events by which oxacene 1 is
transformed to 5 could not be readily ascertained as multiple
independent pathways can be envisioned for the sequential
Staudinger/aza-Claisen-retro-Claisen process, all of which
afford azocine 5 after hydrolysis (Scheme 4). Use of excess
Staudinger reagent apparently drives the formation of the
diimine 10 or the azocine imine 11 ensuring complete
conversion to 5 after hydrolysis.
Condensation of aldehyde alcohol 3 with benzyl amine
afforded aldimine alcohol 4. Again 4 did not undergo aza-
retro-Claisen rearrangement until after oxidation of 4 with
Dess-Martin periodinane, which afforded the azocine al-
dehyde 5 directly at rt with no detectable cyclobutane
aldehyde imine by NMR (Keq >20). Although the shift in
the equilibrium constant for interconversion of 4 and 5 owing
to the combined effects of ring strain relief and additional
resonance stabilization was expected, the apparent reduction
in the kinetic barrier for interconversion of 4 and 5 relative
to interconversion of 2 and 3 was not.
With the feasibility of the aza-retro-Claisen rearrangement
via a stepwise sequence established, we turned our efforts
to development of a one-pot procedure to avoid need for
additional redox transformations.
It has long been established that the equilibrium of the
Claisen/retro-Claisen rearrangement can be perturbed by the
introduction of electron-withdrawing/-donating substitu-
ents.8a,10 By taking advantage of this phenomenon, we
anticipated that conversion of 6 to imine 7 via condensation
with a primary amine or amine equivalent would allow
To further test the scope and functional group compatibility
of the aza-retro-Claisen rearrangement, a series of variously
substituted vinyl cyclobutanecarboxaldehydes/oxacenes were
(11) All materials employed in these studies were racemic although we
have previously described routes to such enantiomerically enriched starting
materials.8
(12) (a) Staudinger, H.; Meyer, J. HelV. Chim. Acta 1919, 2, 635–646.
(b) Staudinger, H.; Hauser, E. HelV. Chim. Acta 1921, 4, 861–886.
(13) Ono, M.; Tamura, S. Chem. Pharm. Bull. 1982, 30, 1453–1462.
(9) We described a single example of azepine formation by rearrange-
ment of the N-benzyl imine of (1S,2S)-1-hydroxymethyl-2-((E)-propen-1-
yl)cyclopropane carbaldehyde after oxidation.8b
(10) Rhoads, S. J.; Cockroft, R. D. J. Am. Chem. Soc. 1969, 91, 2815–
2816
.
Org. Lett., Vol. 12, No. 7, 2010
1629

Scheme 4
.
Potential One-Pot Aza-Retro-Claisen Rearrangement
Table 1. Evaluation of Scope: Azide Component
Pathways to Azocine 5
prepared using variations of our previously described general
route (Scheme 5).8c,11 The sequences were designed to rapidly
access a diversity of substitution rather than providing an optimal
route to any individual example. Beginning with allylic carbonates
12, these sequences afforded the cyclobutanecarboxaldehydes 14
and 15 and/or the oxacenes 16 and 17 in 15-53% combined
overall yield (∼70-80%/step unoptimized) depending upon the
magnitude of Keq for the individual cases at rt.
a For entries 1, 3, and 5 the intermediate formyl imines were hydrolyzed
with pH ) 4 NaOAc buffer. b Yields based on isolated products after
purification by chromatography. c Isolated as 24, the N-H azocine.
N-substituted azocine products 5, 18, and 19 (EWG ) CHO)
and 20-23 (EWG ) SO2Ph) in 75-91% yield. Particularly
noteworthy is the successful use of TMSN3 as a Staudinger
precursor.15 Typically removal of N-protecting groups such as
that present in 19 (mild base) proceeds readily.16 Desilylation
of 22 (N-TMS) occurs during chromatographic purification
affording 24 (N-H), providing ready access to N-unsubstituted
azocines, if desired, minimizing the need for protection/
deprotection steps.
Scheme 5. General Sequence for Preparation of
Cyclobutanecarboxaldehyde/Oxacene Substrates
We also examined the olefin substitution pattern for the vinyl
cyclobutanecarboxaldehydes. For these studies, we employed
a sulfone as the electron-withdrawing group (EWG), eliminating
the need for the acid hydrolysis required when employing a
formyl group as the EWG. The unsubstituted azocine 25 was
readily accessible as depicted in Table 2. Methyl substitution
at the C7 and C8 positions of the oxacene, respectively, gave
products 26 and 20 in excellent yield. Disubstitution was also
well-tolerated, affording the dimethyl azocine 27 in 92% yield,
prompting us to explore the possibility of producing bicyclic
azocines from an appropriate cycloalkenyl-substituted cyclobu-
tane. To our delight the rearrangement afforded the desired
[5,8]-bicyclic azocine 28 in 76% yield. A silyl enol ether was
also employed to determine whether heteroatom substitution
was also tolerated under the standard reaction conditions. As
hoped, the rearrangement proceeded smoothly to afford the
cyclic enol ether 29 in 88% yield. As would be expected, the
one-pot aza-retro-Claisen sequence readily affords azepines from
Having demonstrated the feasibility of a one-pot sequence
to form azocine 5, the scope of nitrogen substitution was
investigated (Table 1). A variety of azides14 were converted to
the respective Staudinger reagents in situ and reacted with
oxacenes 1 and 14 (EWG ) SO2Ph, R1 ) CH3, R2 ) H). As
shown in Table 1, the scope of the nitrogen substitution proved
to be broad. Benzylic, primary, and functionalized primary alkyl,
aryl, and trialkylsilyl are tolerated, affording the expected
(15) Tsuge, O.; Kanemasa, S.; Matsuda, K. J. Org. Chem. 1984, 49,
2688–2691.
(16) Boeckman, R. K., Jr; Jackson, P. F.; Sabatucci, J. P. J. Am. Chem.
Soc. 1985, 107, 2191–2192.
(14) Alvarez, M. T.; Miho, T. Synthesis 1997, 4, 413–414.
1630
Org. Lett., Vol. 12, No. 7, 2010

retroClaisen equilibrium apparently strongly favors oxacen-
ecarboxaldehye 32 over the vinylcyclobutane 31,17 exposure
to 2 equiv of a Staudinger reagent consumed only 1 equiv,
affording the oxacene 33, which failed to undergo the
subsequent Claisen/retro-Claisen rearrangement intended to
provide the desired gem-dimethyl-substituted azocine 34.
Table 2. Evaluation of Scope: Olefin Substitution
In the case of vinylcyclopropane/oxepine 36a/35a and
cyclobutanecarboxaldehyde/oxacene 36b/35b bearing a sul-
fone as the EWG, the equilibrium lies completely (>95%
by NMR) toward oxepine 35a and vinylcyclobutane 36b at
rt. However, upon treatment of 35a/35b and 36a/36b with
excess benzyl Staudinger reagent, the only products observed
were 37a and 37b with the latter isolated in 71% yield
(Scheme 7). These results suggest that the presence of the
combination of syn geminal and nitrogen substituents in 37a
and 37b, thermodynamically disfavors the aza-retro-Claisen
rearrangement even at elevated temperatures.
Scheme 7a
a All reactions employed 3 equiv of Staudinger reagent at 0.2 M in
PhCH3 at reflux for 48 h.
In summary, we have described a practical method for
preparing medium sized N-heterocycles through a one-pot
aza-retro-Claisen rearrangement sequence. The rearrangement
sequence proceeds under mild conditions and appears quite
general and compatible with a variety of substituents at
nitrogen and carbon.
a Yields based on isolated products after chromatography. b For entry
7, related oxepine sulfone used as starting material. c For entry 7, phenyl
azide used in preparation of the Staudinger reagent.
the respective vinylcyclopropanecarboxaldeydes, for example,
providing 30 in 97% yield.
Acknowledgment. We thank Novartis Pharmaceuticals for
partial support of this work and the University of Rochester
(T.R.R. and N.E.G.) and the U.S. Dept. of Education
(GAANN, N.E.G.) for fellowships.
One limitation in the scope of the aza-retro-Claisen
rearrangement encountered thus far involves use of terminally
disubstituted olefins (Scheme 6). Although the Claisen/
Supporting Information Available: Full characterization
data and NMR spectra of all new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
Scheme 6a
OL100397Q
(17) Reactions heated at reflux in PhMe for 72 h with no change.
The temperature was then increased to 180 °C. Heating for 18 h using
a sealed tube afforded 33 with some slight decomposition by 1H
NMR.
a All reactions employed 3 equiv of Staudinger reagent at 0.2 M
concentration in PhCH3 at reflux for 48 h.
Org. Lett., Vol. 12, No. 7, 2010
1631